My blog of building a quiet mini-itx computer
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Re: My blog of building a quiet mini-itx computer
In the meantime I found this site
http://www.coolingtechnique.com/recensi ... eobus.html
Loads of reviews on fans, and most of them I'm interested in.
http://www.coolingtechnique.com/recensi ... eobus.html
Loads of reviews on fans, and most of them I'm interested in.
Re: My blog of building a quiet mini-itx computer
The Scythe PWM A has arrived and I've done some testing with the fan at the highest setting and the lowest setting.
In order to stop the motherboard wanting to speed up the fan when the cpu gets hot, I've used the Zalman fanmate to vary the voltage.
So the Scythe PMW A is not as good as the Scythe 13 as a front fan or the Noctua as a front fan.
Comparing all fans at max on Intelburn, when the Scythe PWM A is blowing at 1980, it is actually not much better than when it is at 1350 (77.5dba vs 65.5dba)
Also comparing all fans at max on Intelburn, the Scythe PWM A at 1980 is no better than a Noctua at 1115.
I guess all the air is going in but not coming out fast enough.
I also used one of the Noctua's black resistors to bring the Scythe PWM A down to about 440 rpm and measure the sound and temps - not better than the Scythe 13 or Noctua as a front fan.
Also using the Noctua's black resistor, the rpm range of the Scythe PWM A went from 1040-1980 to 430-770 !
In order to stop the motherboard wanting to speed up the fan when the cpu gets hot, I've used the Zalman fanmate to vary the voltage.
Code: Select all
Amb.dba 54
Program Idle Idle Intelburn Intelburn Intelburn Intelburn Intelburn
Fans All idle All max All max cpu max top max front max All min
Top fan Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li
rpm 560 1140 1140 560 1140 560 560
dba 59.5 66.5 66.5 60 63.5 64.5 59.5
Front fan Scythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm a
rpm 1040 1980 1980 1040 1040 1980 1040
dba 61.5 77.5 77.5 61.5 61.5 77.5 61.5
Cpu fan Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13
rpm 440 1050 1050 1050 440 440 440
Side dba 57 67 67 57.5 57.5 67 57
Back dba 56.5 62.5 62.5 57 57.5 62.5 56.5
Back PSU dba 57.5 61.5 61.5 57.5 58 62 57.5
Core 0 44 35 72 78 86 85 91
Core 1 44 34 82 86 96 95 99
Core 2 44 33 83 87 93 93 99
Core 3 46 40 78 83 91 91 94
Memory Ambient 30 30 30 30 30 30 30
Cpu package 45 35 82 86 95 95 99
Cpu IA cores 45 35 82 86 95 95 99
Cpu GT cores 35 28 56 61 68 67 73
Motherboard 28 26->24 24->29 26->34 28->30 27->30 26->31
Cpu 35 24->20 21->72 23->77 28->86 24->85 25->89
Pch 49 46 46->47 46->55 50->58 46->52 46->60
HD Assembly 24 22->19 19->22 21->27 24->26 20->24 20->25
HD Airflow 22 20 20 21->22 20 20 20
GPU 29 26->25 25->27 26->33 27->31 25->27 26->29
Amb.dba 54
Program Idle Idle Intelburn Intelburn Intelburn Intelburn Intelburn Intelburn
Fans All idle All max All max cpu max top max front max All min cpu max
Top fan Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li
rpm 560 1140 1140 560 1140 560 560 560
dba 59.5 63.5 63.5 59.5 63.5 60.5 59.5 59.5
Front fan Scythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm a
rpm 550 1350 1350 550 550 1350 550 440
dba 60.5 65.5 65.5 60.5 60.5 65.5 60.5 60.5
Cpu fan Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13 Scythe 13
rpm 440 1050 1050 1050 440 440 440 1050
Side dba 56.5 58.5 58.5 56.5 56.5 58.5 56.5 56.5
Back dba 56.5 58.5 58.5 56.5 57 57.5 56.5 56.5
Back PSU dba 57 59 59 58 57.5 58.5 57 57
Core 0 44 32 73 79 85 89 93 80
Core 1 44 32 81 86 94 99 102 87
Core 2 45 34 81 86 93 97 101 88
Core 3 47 33 78 84 90 93 97 86
Memory Ambient 30 30 30 30 30 30 30 30
Cpu package 45 34 81 86 95 99 102 88
Cpu IA cores 45 34 81 86 95 99 102 87
Cpu GT cores 36 28 55 61 67 29->72 27->75 63
Motherboard 24->31 24 24->29 25->38 31->34 27->31 23->36 31->40
Cpu 25->35 19->23 20->71 20->76 26->84 23->89 21->92 29->78
Pch 46->52 21->22 46->49 46->55 46->60 46->57 46->64 47->57
HD Assembly 20->26 21->22 20->23 20->29 27->28 23->24 21->29 26->31
HD Airflow 21->25 25 24->25 24 24 24 23 23->25
GPU 25->29 25->26 25->27 26->35 28->32 26->28 27->32 29->36
Comparing all fans at max on Intelburn, when the Scythe PWM A is blowing at 1980, it is actually not much better than when it is at 1350 (77.5dba vs 65.5dba)
Also comparing all fans at max on Intelburn, the Scythe PWM A at 1980 is no better than a Noctua at 1115.
I guess all the air is going in but not coming out fast enough.
I also used one of the Noctua's black resistors to bring the Scythe PWM A down to about 440 rpm and measure the sound and temps - not better than the Scythe 13 or Noctua as a front fan.
Also using the Noctua's black resistor, the rpm range of the Scythe PWM A went from 1040-1980 to 430-770 !
Re: My blog of building a quiet mini-itx computer
I'm looking at the cooltechniques fan results against the results I'm getting when the fans are in a pc and try to understand why?
I've typed in the raw numbers into a spreadsheet to get my own comparatives.
Because the spreadsheet is for my own purposes, I've had to cut off the fans' results at 5V for non-pwm fans.
It looks like for a cpu fan that is buried inside the pc, anything below 12dba on the cooltechnique website will be inaudible.
But the performance of the Scythe 13 versus the Noctua as a front fan or cpu fan is something I can't explain.
I think because of the combination of grills and fins inside the Lian-Li case, the dba vs pressure is more important than dba vs cfm.
The coolink 120P is quieter with better pressure than the Scythe 13 at minimum rpm, but loses to the Scythe as soon as the rpm's get higher.
But as you can see from the chart, for dba vs pressure the Scythe 13 is one of the best.
The Coolink 1201 goes down to 3V, has the same pressure at the Scythe 13 and much lower noise. But I can't get down to 3V with the Zalman fanmate controller. Hmmm, are there any other fan controllers going down to 3V ... off to google.
I've typed in the raw numbers into a spreadsheet to get my own comparatives.
Because the spreadsheet is for my own purposes, I've had to cut off the fans' results at 5V for non-pwm fans.
It looks like for a cpu fan that is buried inside the pc, anything below 12dba on the cooltechnique website will be inaudible.
But the performance of the Scythe 13 versus the Noctua as a front fan or cpu fan is something I can't explain.
I think because of the combination of grills and fins inside the Lian-Li case, the dba vs pressure is more important than dba vs cfm.
The coolink 120P is quieter with better pressure than the Scythe 13 at minimum rpm, but loses to the Scythe as soon as the rpm's get higher.
But as you can see from the chart, for dba vs pressure the Scythe 13 is one of the best.
The Coolink 1201 goes down to 3V, has the same pressure at the Scythe 13 and much lower noise. But I can't get down to 3V with the Zalman fanmate controller. Hmmm, are there any other fan controllers going down to 3V ... off to google.
You do not have the required permissions to view the files attached to this post.
Re: My blog of building a quiet mini-itx computer
OK - I've found this thread where someone is making pwm to analogue controllers. I don't have space on the Lian Li PC-Q08 for a fancy fan controller but this should work just as well.
http://forums.overclockers.co.uk/showth ... t=18421145
So 3V fans are back on the list.
I've added some more values to the charts from cooltechique. It's getting crowded in there!
Looking at the charts again, I'm convinced that pressure is more important than cfm.
From here
http://www.xbitlabs.com/picture/?src=/i ... 33_xbt.png
The Noiseblocker Multiframe S2 looks good but from the chart I've produced from cooltechnique, it agrees on cfm (Noiseblocker better than Coolink)
but I look for the lowest noise, then the Coolink beats the Noiseblocker on dba vs pressure.
The pressure chart partly explains why at the same noise level, the Scythe 13 can beat the Noctua P14, it has pressure of about 1.0 versus the Noctua pressure of about 0.5 at around the 12.5dba level (for the cpu fan performance). but I don't know why the Scythe 14 came out so badly ... a bad sample? The Scythe Glidestream noise and pressure levels follow on quite smoothly from each other as you increase speed, so you would expect the same for the Scythe Gentle Typhoons.
For the top and front fan performance, when the fan is next to the opening (with a grill/filter), then the Coolinks gives a similar pressure to the Scythe 13 but at quieter noise levels.
But it also looks like the Coolink will be noisier than the Scythe at full speed and for not much better performance at the top end.
Back to have another look at the spreadsheets!
http://forums.overclockers.co.uk/showth ... t=18421145
So 3V fans are back on the list.
I've added some more values to the charts from cooltechique. It's getting crowded in there!
Looking at the charts again, I'm convinced that pressure is more important than cfm.
From here
http://www.xbitlabs.com/picture/?src=/i ... 33_xbt.png
The Noiseblocker Multiframe S2 looks good but from the chart I've produced from cooltechnique, it agrees on cfm (Noiseblocker better than Coolink)
but I look for the lowest noise, then the Coolink beats the Noiseblocker on dba vs pressure.
The pressure chart partly explains why at the same noise level, the Scythe 13 can beat the Noctua P14, it has pressure of about 1.0 versus the Noctua pressure of about 0.5 at around the 12.5dba level (for the cpu fan performance). but I don't know why the Scythe 14 came out so badly ... a bad sample? The Scythe Glidestream noise and pressure levels follow on quite smoothly from each other as you increase speed, so you would expect the same for the Scythe Gentle Typhoons.
For the top and front fan performance, when the fan is next to the opening (with a grill/filter), then the Coolinks gives a similar pressure to the Scythe 13 but at quieter noise levels.
But it also looks like the Coolink will be noisier than the Scythe at full speed and for not much better performance at the top end.
Back to have another look at the spreadsheets!
You do not have the required permissions to view the files attached to this post.
Re: My blog of building a quiet mini-itx computer
Some more results
I did a test with the Scythe pwm a turned up as much as possible but keeping the top noise level to 59.5 and sides to 56.5. the lowest I have so far, and matches the Scythe 13.
The Scythe pwm a rpm went up to about 790 whereas the Scythe 13 went to 1050.
The cpu temp with the Scythe pwm a was about 5C worse than the Scythe 13 for the same noise level. That's a big giveaway.
Now I'm thinking the Scythe 14 might be a good choice for the cpu fan if it performs similarly to the Scythe 13 but at a higher rpm.
OK -had another read of the silentpcreview of the 1450 - looks like there might be a tonal ring at less than 1100rpm, but that is about the speed I will be running it at anyway, and amongst the other noise, I hope I won't hear it. I don't hear any such noise with the Scythe 13 I have at the moment.
Code: Select all
Program Idle Idle Intelburn Intelburn Intelburn Intelburn Intelburn
Fans All idle All max All max cpu max top max front max All min
Top fan Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li
rpm 560 1140 1140 560 1140 560 560
dba 72.5 72.5 64.5 60.5 60.5
Front fan Noctua Noctua Noctua Noctua Noctua 62.5 Noctua
rpm 475 1115 1115 475 475 1115 475
dba 69.5 69.5 60.5 62.5 60
Cpu fan Scythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm a
rpm 950 1875 1875 1875 950 950 950
Side dba 66.5 65.5 58 58 57.5
Back dba 70 59.5 58.5 58.5 57.5
Back PSU dba 68.5 68.5 58.5 59.5 58.5
Core 0 37 34 68 74 74 78 81
Core 1 33 32 78 82 83 86 87
Core 2 35 29 76 82 82 85 87
Core 3 40 35 76 80 80 83 85
Memory Ambient 30 30 30 80 30 30 30
Cpu package 40 34 78 82 83 85 88
Cpu IA cores 40 34 78 82 83 85 88
Cpu GT cores 32 27 51 57 57 59 62
Motherboard 30->31 24->25 24->30 25->38 26->33 26->31 23->37
Cpu 23->28 19->22 21->68 22->72 21->73 22->76 23->78
Pch 46->47 46 46 46->50 46->52 46->52 46->58
HD Assembly 27->28 20->21 20->24 21->29 22->26 23->25 23->30
HD Airflow 23->24 22->23 22 22 22->23 23 23
GPU 30->32 25->26 25->29 27->38 27->30 27->29 28->34
Program Idle Idle Intelburn Intelburn Intelburn Intelburn Intelburn Intelburn
Fans All idle All max All max cpu max top max front max All min low noise
Top fan Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li Lian Li
rpm 560 1140 1140 560 1140 560 560 560
dba 59.5 65.5 65.5 62.5 63.5 60.5 59.5 59.5
Front fan Noctua Noctua Noctua Noctua Noctua 62.5 Noctua Noctua
rpm 475 1115 1115 475 475 1115 475 475
dba 59.5 63.5 63.5 61.5 60.5 62 59.5 59.5
Cpu fan Scythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm aScythe pwm a
rpm 445 1215 1215 1215 445 445 445 790
Side dba 56.5 59.5 59.5 58.5 57.5 58 56.5 56.5
Back dba 56.5 59.5 59.5 59.5 57.5 57.5 56.5 56.5
Back PSU dba 57 61 61 60.5 58.75 58.5 57 57.5
Core 0 39 33 71 77 85 85 94 82
Core 1 40 35 81 85 95 95 102 93
Core 2 43 38 79 86 93 93 100 90
Core 3 47 39 78 82 90 90 96 88
Memory Ambient 30 30 30 30 30 30 30 30
Cpu package 44 38 80 86 95 96 102 93
Cpu IA cores 44 38 80 86 95 96 102 93
Cpu GT cores 36 32 54 59 68 68 74 65
Motherboard 29 23->24 24->29 28->38 31->33 23->30 26->34 28->37
Cpu 25->29 19->24 21->71 22->76 26->85 21->85 23->92 23->82
Pch 48->50 46 46->47 46->55 46->60 40->54 46->63 46->60
HD Assembly 25->26 20->21 21->23 23->30 28->29 20->23 22->27 23->30
HD Airflow 25 24->25 24 24->25 25 24 24 24
GPU 28->29 25 25->28 29->36 28->33 25->27 26->30 28->33
The Scythe pwm a rpm went up to about 790 whereas the Scythe 13 went to 1050.
The cpu temp with the Scythe pwm a was about 5C worse than the Scythe 13 for the same noise level. That's a big giveaway.
Now I'm thinking the Scythe 14 might be a good choice for the cpu fan if it performs similarly to the Scythe 13 but at a higher rpm.
OK -had another read of the silentpcreview of the 1450 - looks like there might be a tonal ring at less than 1100rpm, but that is about the speed I will be running it at anyway, and amongst the other noise, I hope I won't hear it. I don't hear any such noise with the Scythe 13 I have at the moment.
Re: My blog of building a quiet mini-itx computer
Another update.
No more spreadsheet results for the moment as I think the results are too specific to my particular configuration.
But I did compare ducting with no ducting.
Core temps might be 1-2C higher with ducting unless it is at min rpm for all the fans in which case it is better by about 7-9C! This is well worth having at min rpm.
Also the HD appears to be 1-2C cooler; I assume from not having the hot air from the cpu being blown back to the HD area.
Looking at my results, I think it is a 'balance' of airflow which makes the difference when there is ducting.
Each of the cpu, top and front fan could be a limiting factor for the other fans
So there is no point in increasing the speed of one fan without the others.
That would explain why at min rpm, there is not much difference in increasing each of the other fans (except the cpu fan)
Overall, I think the ducting is a good thing.
The Coolink 1201 arrived and I did some testing
As a front fan the Coolink is a disappointment.
Even at max rpm, it cannot match the Noctua at lowest rpm for cooling power.
And the Scythe 13 matched the Noctua for cooling power at 1 dba less.
Disappointed.
This changes when the Coolink is used as a topfan.
The Coolink doesn't sound quieter than the Lian Li but the dba says it is 1db quieter.
I then put the Noctua black resistor in line and bought the Coolink from the Zalman 5V (650rpm) down to 450rpm.
Now the dba meter is saying the Coolink is at 56.5dba, the quiestest measurement so far for a top fan.
I swap the Scythe 13 back into the top fan position and it reads 59.5.
But is doesn't sound louder than the Coolink.
After doing further testing, the Scythe at min rpm is about 5C cooler than Coolink as a topfan at not much louder noise to my ear.
While hunting around for Scythe Gentle Typhoon reviews I came across this post
http://www.overclock.net/t/1273923/coug ... e-begin/30
where Martinm210 says "The AP-15s do have issues with harmonics at specific RPM levels, but near as I can tell...nothing comes even remotely close in keeping noise levels as low when rad mounted."
So basically GT are very good when there is a resistance.
So I've ordered a Scythe GT AP14 to go in as a cpu fan cooler.
But further on in response to a question about bearing noise
"Get a syringe or poke a hole in the rear label and squirt some tri-lube in there."
which led me to here
http://www.overclockers.com/em-spinning ... e-pc-fans/
So I've ordered some sewing machine oil to see if I can eliminate that bearing noise from the Scythe AP-15 I had. I might try that with the other fans I've got
Hmmm ... I wonder how quiet they might become ...
No more spreadsheet results for the moment as I think the results are too specific to my particular configuration.
But I did compare ducting with no ducting.
Core temps might be 1-2C higher with ducting unless it is at min rpm for all the fans in which case it is better by about 7-9C! This is well worth having at min rpm.
Also the HD appears to be 1-2C cooler; I assume from not having the hot air from the cpu being blown back to the HD area.
Looking at my results, I think it is a 'balance' of airflow which makes the difference when there is ducting.
Each of the cpu, top and front fan could be a limiting factor for the other fans
So there is no point in increasing the speed of one fan without the others.
That would explain why at min rpm, there is not much difference in increasing each of the other fans (except the cpu fan)
Overall, I think the ducting is a good thing.
The Coolink 1201 arrived and I did some testing
As a front fan the Coolink is a disappointment.
Even at max rpm, it cannot match the Noctua at lowest rpm for cooling power.
And the Scythe 13 matched the Noctua for cooling power at 1 dba less.
Disappointed.
This changes when the Coolink is used as a topfan.
The Coolink doesn't sound quieter than the Lian Li but the dba says it is 1db quieter.
I then put the Noctua black resistor in line and bought the Coolink from the Zalman 5V (650rpm) down to 450rpm.
Now the dba meter is saying the Coolink is at 56.5dba, the quiestest measurement so far for a top fan.
I swap the Scythe 13 back into the top fan position and it reads 59.5.
But is doesn't sound louder than the Coolink.
After doing further testing, the Scythe at min rpm is about 5C cooler than Coolink as a topfan at not much louder noise to my ear.
While hunting around for Scythe Gentle Typhoon reviews I came across this post
http://www.overclock.net/t/1273923/coug ... e-begin/30
where Martinm210 says "The AP-15s do have issues with harmonics at specific RPM levels, but near as I can tell...nothing comes even remotely close in keeping noise levels as low when rad mounted."
So basically GT are very good when there is a resistance.
So I've ordered a Scythe GT AP14 to go in as a cpu fan cooler.
But further on in response to a question about bearing noise
"Get a syringe or poke a hole in the rear label and squirt some tri-lube in there."
which led me to here
http://www.overclockers.com/em-spinning ... e-pc-fans/
So I've ordered some sewing machine oil to see if I can eliminate that bearing noise from the Scythe AP-15 I had. I might try that with the other fans I've got
Hmmm ... I wonder how quiet they might become ...
Re: My blog of building a quiet mini-itx computer
From here
http://www.youtube.com/watch?v=eelgZSv_Hh8
"The gentle typhoons are all the same fan just modified with resistors to spin at different speeds."
Does that mean all the differences in Gentle Typhoons noise at the same rpm is due to lubricant levels between different fans rather than different motors inside each one?
That would explain why the Scythe AP-15 has such good results in this review
http://www.xbitlabs.com/articles/cooler ... html#sect0
And all other variations are down to the amount of lubricant in the sample.
http://www.youtube.com/watch?v=eelgZSv_Hh8
"The gentle typhoons are all the same fan just modified with resistors to spin at different speeds."
Does that mean all the differences in Gentle Typhoons noise at the same rpm is due to lubricant levels between different fans rather than different motors inside each one?
That would explain why the Scythe AP-15 has such good results in this review
http://www.xbitlabs.com/articles/cooler ... html#sect0
And all other variations are down to the amount of lubricant in the sample.
Re: My blog of building a quiet mini-itx computer
I've been gathering loads of readings and I've had a chance to put them together to see what might be the best fans for my particular build.
One interesting result is that under IntelBurn, the lowest "cpu package" temperature I can get is 78C. I can reach that 'low' with all the fans at full blast (various different combinations) at high noise levels, or I also managed to reach that with Scythe 13 top fan @1050 rpm, Noctua P14 front fan @ 480 rpm and Noctua P14 cpu fan @ 800rpm, which produces very little sound. Speeding up the cpu fan and/or front fan doesn't seem to significantly reduce the temperature. That's kind of interesting.
Still running some more tests and waiting for the oil to arrive.
One interesting result is that under IntelBurn, the lowest "cpu package" temperature I can get is 78C. I can reach that 'low' with all the fans at full blast (various different combinations) at high noise levels, or I also managed to reach that with Scythe 13 top fan @1050 rpm, Noctua P14 front fan @ 480 rpm and Noctua P14 cpu fan @ 800rpm, which produces very little sound. Speeding up the cpu fan and/or front fan doesn't seem to significantly reduce the temperature. That's kind of interesting.
Still running some more tests and waiting for the oil to arrive.
Re: My blog of building a quiet mini-itx computer
Not good news
I applied the oil to the Scythe 13 & Scythe 15 bearings and they became noticeably noisier.
At full rpm at 0cm, they have become 4-5 dba noisier after lubricating them.
I applied the oil to the Scythe 13 & Scythe 15 bearings and they became noticeably noisier.
At full rpm at 0cm, they have become 4-5 dba noisier after lubricating them.
Re: My blog of building a quiet mini-itx computer
It hasn't worked for me.
The Scythe AP-15 mounted as top fan of Lian-Li PC-Q08 case, at 0cm.
At 750rpm, it was originally 65.5dba, after three drops of sewing oil it was 69 dba, after three drops of 10-40w oil it is 71dba
At 1600rpm, it was originally 73dba, after sewing oil it was 79 dba, after 10-40w oil it is 81dba
As a comparison, the Scythe AP-13 was originally measuring 61dba @460rpm and 65.5dba @ 1050rpm, and I was trying to make the Scythe AP-15 match the AP-13 sound levels.
The Scythe AP-15 mounted as top fan of Lian-Li PC-Q08 case, at 0cm.
At 750rpm, it was originally 65.5dba, after three drops of sewing oil it was 69 dba, after three drops of 10-40w oil it is 71dba
At 1600rpm, it was originally 73dba, after sewing oil it was 79 dba, after 10-40w oil it is 81dba
As a comparison, the Scythe AP-13 was originally measuring 61dba @460rpm and 65.5dba @ 1050rpm, and I was trying to make the Scythe AP-15 match the AP-13 sound levels.
Re: My blog of building a quiet mini-itx computer
A discovery from all my testing.
The front fan cools the hard disk and graphics card but is not critical to the cpu temperature.
The cpu fan and the top fan are critical to the cpu temperature with the cpu fan being more important.
Because of the way my build is set up, the Noctua NH-C14 pulls half of its air from within the case after it has passed through the HD or from the graphics card, but it pulls the other half of its air from the outside via the Lazer LZP-500 PSU which doesn't spin up. And pulling the air past the PSU cools it down anyway! So although the PSU is directly above the heatsink and would normally block it, because the intake of the PSU is facing the heatsink and not the outside of the Lian-Li case, it provides an air intake for the heatsink. Brilliant!
The front fan cools the hard disk and graphics card but is not critical to the cpu temperature.
The cpu fan and the top fan are critical to the cpu temperature with the cpu fan being more important.
Because of the way my build is set up, the Noctua NH-C14 pulls half of its air from within the case after it has passed through the HD or from the graphics card, but it pulls the other half of its air from the outside via the Lazer LZP-500 PSU which doesn't spin up. And pulling the air past the PSU cools it down anyway! So although the PSU is directly above the heatsink and would normally block it, because the intake of the PSU is facing the heatsink and not the outside of the Lian-Li case, it provides an air intake for the heatsink. Brilliant!
Re: My blog of building a quiet mini-itx computer
I spent some time over Xmas putting in the butyl sound deadening material and then putting the acoustic foam 4mm and acoustic absorber 15mm into the case.
Once I had finished putting the 2mm sound deadening material on the case, I realised there was space in the hard drive supports, cages etc. So I then used up more sound deadening material on that. The case is a lot heavier now and thuds rather than rings.
I then started to put on the 4mm and 15 mm foam ... and found out that actually, there was much more 'space' inside the Q08 than I thought, so I've run out of the 15mm foam and I've had to order some more ... waiting for that to arrive.
In the meantime, while tapping on the case, I found that when I tapped the top of the case, there was a ringing sound. I tracked it down to the CD drive. ... just waiting for the post now.
Once I had finished putting the 2mm sound deadening material on the case, I realised there was space in the hard drive supports, cages etc. So I then used up more sound deadening material on that. The case is a lot heavier now and thuds rather than rings.
I then started to put on the 4mm and 15 mm foam ... and found out that actually, there was much more 'space' inside the Q08 than I thought, so I've run out of the 15mm foam and I've had to order some more ... waiting for that to arrive.
In the meantime, while tapping on the case, I found that when I tapped the top of the case, there was a ringing sound. I tracked it down to the CD drive. ... just waiting for the post now.
Re: My blog of building a quiet mini-itx computer
Strange - I've tried posting some pictures but the board won't let me 800x600, 100kb or so.
Re: My blog of building a quiet mini-itx computer
Some pictures of the final build.
The inside
The other side of the inside
Asus P8Z77-I with Noctua NH-C14 - It can be done!
The other side
The inside
The other side of the inside
Asus P8Z77-I with Noctua NH-C14 - It can be done!
The other side
Re: My blog of building a quiet mini-itx computer
More pictures
The fan
The fan holder
Asus P8Z77-I with Noctua NH-C14 inside Lian-Li PC-Q08
Asus P8Z77-I with Noctua NH-C14 inside Lian-Li PC-Q08 and Lazerwin LZP-550 installed.
The wires look really bad but what you can't see from the picture is that there is 'space' underneath the wiring for airflow. The wiring will rest against the outside edge of the case leaving airflow inside the case.
The fan
The fan holder
Asus P8Z77-I with Noctua NH-C14 inside Lian-Li PC-Q08
Asus P8Z77-I with Noctua NH-C14 inside Lian-Li PC-Q08 and Lazerwin LZP-550 installed.
The wires look really bad but what you can't see from the picture is that there is 'space' underneath the wiring for airflow. The wiring will rest against the outside edge of the case leaving airflow inside the case.
Re: My blog of building a quiet mini-itx computer
Initial impression were the hard disk was a lot quieter, about half as loud as before ... but I didn't measure the hard disk noise Less vibration felt in the case. Definitely no rattling or ringing from the case.
Disappointment with the front and top fan noise, no difference at all. But also, not much change with the temperatures either, so all the sound dampening material I put in hadn't interfered with the airflow.
The DVD is quieter. It's basically covered in sound dampening material.
Cpu fan noise is muted a little bit but that was quite quiet to begin with.
Disappointment with the front and top fan noise, no difference at all. But also, not much change with the temperatures either, so all the sound dampening material I put in hadn't interfered with the airflow.
The DVD is quieter. It's basically covered in sound dampening material.
Cpu fan noise is muted a little bit but that was quite quiet to begin with.
Re: My blog of building a quiet mini-itx computer
Some charts.
Temperature of CPU Package by HWInfo, by top fan speed for different cpu fan speeds (Both top and cpu fans being Scythe 15)
I did get cpu throttling on 2 cores on the last measurement of cpu fan @ 355 rpm and top fan @ 380 rpm
Cpu is 3570K overclocked to 4.2GHz
Measurements were taken over several days with room temperature varying between 19C to 22C. The graph is what I get after adjusting for temperature to 21C. Obviously there are errors in measurement but I think there are overall trends in the chart which are valid.
So with the cpu fan @ 355 rpm, any increase in the top fan produces a steady decrease in cpu temp.
With cpu fan @ 1050rpm, and further increase in cpu fan speed does decrease the cpu temp but not as great as before 1050 rpm.
Temperature of cpu by cpu fan speed for different top fan speeds. Same data as above sorted the other way.
So not too much difference in cpu temp when the top fan speed is between 380 and 500 rpm, although there is some.
Once again, most of the cpu cooling gains come at 1050 rpm for the top fan speed. After that, there is improvement but slow going.
Sound levels of airflow of the top fan - measured using an ipad app. Once again, obviously there will be errors in measurement, but hopefully there will be useful information we can extract.
[Edit - copied the wrong row in the spreadsheet - It is airflow noise level of the top fan, and not 25cm away from the top fan]
Yep that 1600 looks funny but that's because the dba meter we jumping between several numbers and didn't settle on a single number so I took what look liked to be the range and divided by two. But the top maximum dba number is relatively constant. This shows the impact of the cpu fan noise on the top fan noise.
With the top fan at 380 rpm, the cpu fan noise is not measurable until after 1050 rpm, which is basically what my ear is telling me too.
Sound levels of airflow of the top fan sorted the other way. Oh boy, even more errors of measurement this way!
[Edit - copied the wrong row in the spreadsheet - It is airflow noise level of the top fan, and not 25cm away from the top fan]
But with the top fan at 380 rpm, you don't notice the cpu fan until after 1050rpm, you can see the 1200 cpu rpm start above that collection of points.
That is only one set of charts. Other numbers I have include front and side noise levels at 25cm and 0cm.
The cpu fan and top fan both impact those numbers to a greater or lesser degree. I'll spare you the results of those!
But basically, run the cpu fan faster than the top fan for best cooling/noise result.
Temperature of CPU Package by HWInfo, by top fan speed for different cpu fan speeds (Both top and cpu fans being Scythe 15)
I did get cpu throttling on 2 cores on the last measurement of cpu fan @ 355 rpm and top fan @ 380 rpm
Cpu is 3570K overclocked to 4.2GHz
Measurements were taken over several days with room temperature varying between 19C to 22C. The graph is what I get after adjusting for temperature to 21C. Obviously there are errors in measurement but I think there are overall trends in the chart which are valid.
So with the cpu fan @ 355 rpm, any increase in the top fan produces a steady decrease in cpu temp.
With cpu fan @ 1050rpm, and further increase in cpu fan speed does decrease the cpu temp but not as great as before 1050 rpm.
Temperature of cpu by cpu fan speed for different top fan speeds. Same data as above sorted the other way.
So not too much difference in cpu temp when the top fan speed is between 380 and 500 rpm, although there is some.
Once again, most of the cpu cooling gains come at 1050 rpm for the top fan speed. After that, there is improvement but slow going.
Sound levels of airflow of the top fan - measured using an ipad app. Once again, obviously there will be errors in measurement, but hopefully there will be useful information we can extract.
[Edit - copied the wrong row in the spreadsheet - It is airflow noise level of the top fan, and not 25cm away from the top fan]
Yep that 1600 looks funny but that's because the dba meter we jumping between several numbers and didn't settle on a single number so I took what look liked to be the range and divided by two. But the top maximum dba number is relatively constant. This shows the impact of the cpu fan noise on the top fan noise.
With the top fan at 380 rpm, the cpu fan noise is not measurable until after 1050 rpm, which is basically what my ear is telling me too.
Sound levels of airflow of the top fan sorted the other way. Oh boy, even more errors of measurement this way!
[Edit - copied the wrong row in the spreadsheet - It is airflow noise level of the top fan, and not 25cm away from the top fan]
But with the top fan at 380 rpm, you don't notice the cpu fan until after 1050rpm, you can see the 1200 cpu rpm start above that collection of points.
That is only one set of charts. Other numbers I have include front and side noise levels at 25cm and 0cm.
The cpu fan and top fan both impact those numbers to a greater or lesser degree. I'll spare you the results of those!
But basically, run the cpu fan faster than the top fan for best cooling/noise result.
Re: My blog of building a quiet mini-itx computer
More conclusions.
Start with as quiet as components as you can!
The hard disk is behind the front fan in front of a grill - direct sound path. So the hard disk affects the front fan noise greatly.
With top and cpu Scythe AP-15 fan @ 750rpm and front Scythe AP-14 fan @ 600rpm and 1 hard disk in place the dba at 0cm measured was 59dba
With top and cpu Scythe AP-15 fan @ 750rpm and no hard disk; the front Scythe AP-14 fan @ 300rpm measured 56dba, 500rpm/57dba, 750rpm/58dba.
So until the front fan reaches 750rpm, the fan noise is being covered by the hard disk.
The top fan is right next to the grill so you just can't sound proof it at all and slow speed is the only way to make it quiet.
Luckily for my configuration, the cpu fan is muffled by the components around it so it can spin up to around 750-1050 before you can hear it above the sound of the other components.
To throw in more complication, the Lazerwin LZP-550 has a switch at the back to turn on the fan.
I didn't do a full set of measurements but when it is turned, at the low rpm range of the cpu and top fan, the psu fan decreases the cpu temps by 2-4C, but a the top rpm range of the cpu and top fan, the psu fan increases the cpu temp by about 2C. Could it be that at low speed it actually help pull out the hot air from the heatsink but at high cpu fan speed (1200+rpm), the psu fan then hinders the airflow in?
I started testing the graphics card 7850 using furmark ... its loud !!!
With room base noise at around 54dba, before the side dba at 0cm was 55dba /1200rpm / 20% but on furmark it measured 71.5dba / 4021 rpm / 71% of full fan speed.
In goes the Scythe Musashi to the rescue!
I don't have space for the fan controller knobs on the pci plate so I've ordered a pwm to analogue controller to control the Scythe Musashi fans from the graphics card pwm output directly.
I've also ordered 2 more pwm to analogue fan controllers for the Scythe fans so they can be controlled by the P8Z77-I fan headers and I won't need to have the three Zalman fanmates hanging out of the back!
Start with as quiet as components as you can!
The hard disk is behind the front fan in front of a grill - direct sound path. So the hard disk affects the front fan noise greatly.
With top and cpu Scythe AP-15 fan @ 750rpm and front Scythe AP-14 fan @ 600rpm and 1 hard disk in place the dba at 0cm measured was 59dba
With top and cpu Scythe AP-15 fan @ 750rpm and no hard disk; the front Scythe AP-14 fan @ 300rpm measured 56dba, 500rpm/57dba, 750rpm/58dba.
So until the front fan reaches 750rpm, the fan noise is being covered by the hard disk.
The top fan is right next to the grill so you just can't sound proof it at all and slow speed is the only way to make it quiet.
Luckily for my configuration, the cpu fan is muffled by the components around it so it can spin up to around 750-1050 before you can hear it above the sound of the other components.
To throw in more complication, the Lazerwin LZP-550 has a switch at the back to turn on the fan.
I didn't do a full set of measurements but when it is turned, at the low rpm range of the cpu and top fan, the psu fan decreases the cpu temps by 2-4C, but a the top rpm range of the cpu and top fan, the psu fan increases the cpu temp by about 2C. Could it be that at low speed it actually help pull out the hot air from the heatsink but at high cpu fan speed (1200+rpm), the psu fan then hinders the airflow in?
I started testing the graphics card 7850 using furmark ... its loud !!!
With room base noise at around 54dba, before the side dba at 0cm was 55dba /1200rpm / 20% but on furmark it measured 71.5dba / 4021 rpm / 71% of full fan speed.
In goes the Scythe Musashi to the rescue!
I don't have space for the fan controller knobs on the pci plate so I've ordered a pwm to analogue controller to control the Scythe Musashi fans from the graphics card pwm output directly.
I've also ordered 2 more pwm to analogue fan controllers for the Scythe fans so they can be controlled by the P8Z77-I fan headers and I won't need to have the three Zalman fanmates hanging out of the back!
Re: My woes of building a quiet mini-itx computer
My woes of building a quiet mini-itx computer
The pwm-dc controllers arrived.
The cpu and the dual board top/fan fan controllers work fine.
The only limitation thing is that if you lower the starting rpm, then you also lose a few rpm at the top end.
In between the Asus Bios and Speedfan, I'll sort something out that I will be happy with.
One interesting thing is that at 750/750/450, cpu/top/front fans, the pc was quite able to run intelburn without any problems. The cores were about 87C and the cpu package was about 77C. But when I was running BIONC seti@home, around 77C core and 67C cpu package, it crashed!
I tried the fans at 1050/750/450, ( cpu package average around 59, peak 71C) and the pc didn't crash anymore. Those charts I did earlier is going to useful in deciding where to set the fan speeds!
So, its not the case of how fast I can overclock the pc, but how quietly do I want to run the pc and the fastest speed I can do that quietness.
On another note, the Scythe Musashi fan controllers do something very strange.
When I power the Scythe Mushashi fans from the psu, the fan controllers work fine.
When I use the gpu card pwm connector, connect it to the pwm-dc controller to a 12V fan, the 12V fan works fine.
But when I plug the gpu card pwm connector to the pwm-dc controller to the input of the Scythe Musashi fan, it doesn't work!
The Scythe fans are the Scythe "KAZE JYU SLIM" 100mm case fan with some extra wire from the controller knobs going into it.
But obviously, something is not right.
I've decided to get some normal scythe fans and take off the pci bracket controlled Scythe fans. Those are now on order.
In the meantime, the graphics card decided not to work after I attached the Scythe Musashi heatsink. It got stuck at the 'Starting Windows' screen after booting up. It still works if I uninstall the amd drivers and stick with the Windows default VGA drivers. Windows Performance index of 1.0.
I put the original heatsink and fan back on. No joy. I put the graphics card into my Shuttle and it boots up fine. So, I vacuum the P8Z77-I board to remove any possible dust/fibre that could be causing a problem. Run a chkdsk in case of any problems there. Put the card back in the slot, but no joy.
The pwm-dc controllers arrived.
The cpu and the dual board top/fan fan controllers work fine.
The only limitation thing is that if you lower the starting rpm, then you also lose a few rpm at the top end.
In between the Asus Bios and Speedfan, I'll sort something out that I will be happy with.
One interesting thing is that at 750/750/450, cpu/top/front fans, the pc was quite able to run intelburn without any problems. The cores were about 87C and the cpu package was about 77C. But when I was running BIONC seti@home, around 77C core and 67C cpu package, it crashed!
I tried the fans at 1050/750/450, ( cpu package average around 59, peak 71C) and the pc didn't crash anymore. Those charts I did earlier is going to useful in deciding where to set the fan speeds!
So, its not the case of how fast I can overclock the pc, but how quietly do I want to run the pc and the fastest speed I can do that quietness.
On another note, the Scythe Musashi fan controllers do something very strange.
When I power the Scythe Mushashi fans from the psu, the fan controllers work fine.
When I use the gpu card pwm connector, connect it to the pwm-dc controller to a 12V fan, the 12V fan works fine.
But when I plug the gpu card pwm connector to the pwm-dc controller to the input of the Scythe Musashi fan, it doesn't work!
The Scythe fans are the Scythe "KAZE JYU SLIM" 100mm case fan with some extra wire from the controller knobs going into it.
But obviously, something is not right.
I've decided to get some normal scythe fans and take off the pci bracket controlled Scythe fans. Those are now on order.
In the meantime, the graphics card decided not to work after I attached the Scythe Musashi heatsink. It got stuck at the 'Starting Windows' screen after booting up. It still works if I uninstall the amd drivers and stick with the Windows default VGA drivers. Windows Performance index of 1.0.
I put the original heatsink and fan back on. No joy. I put the graphics card into my Shuttle and it boots up fine. So, I vacuum the P8Z77-I board to remove any possible dust/fibre that could be causing a problem. Run a chkdsk in case of any problems there. Put the card back in the slot, but no joy.
Re: My blog of building a quiet mini-itx computer
Hi,
I'd thought I'd just post a few pictures.
This is a pwm-dc controller for the graphics card.
This the dual board with an extra wire to measure rpm of the B device, and the cpu one.
Before with the Zalman controllers that were hanging out of the back.
Detail of the graphics card before it is covered up. It's a tight fit.
After. It doesn't look neater but it is. There are no more wires coming out of the back.
In the picture you see the Musashi controller wrapped up in plastic to stop it touching anything else. However, as I found out that the Musashi controller is not compatible with the gpu card output and the pwm-dc controller, I've decided to replace the mounted Scythe Musashi fans with the equivalent Scythe fans but without the extra controller circuitry attached. That should then work fine. It also means the pci backplate won't be in the computer and neatens things up a little
I'd thought I'd just post a few pictures.
This is a pwm-dc controller for the graphics card.
This the dual board with an extra wire to measure rpm of the B device, and the cpu one.
Before with the Zalman controllers that were hanging out of the back.
Detail of the graphics card before it is covered up. It's a tight fit.
After. It doesn't look neater but it is. There are no more wires coming out of the back.
In the picture you see the Musashi controller wrapped up in plastic to stop it touching anything else. However, as I found out that the Musashi controller is not compatible with the gpu card output and the pwm-dc controller, I've decided to replace the mounted Scythe Musashi fans with the equivalent Scythe fans but without the extra controller circuitry attached. That should then work fine. It also means the pci backplate won't be in the computer and neatens things up a little
Re: My blog of building a quiet mini-itx computer
Good news in the fact I managed to have the graphics card problem sorted out.
Asus support said to install everything again from a clean disk.
I did uninstall the drivers and deleted the directories used, but that didn't work.
In theory, taking a graphics card out and plugging it back in again should cause any problems. I've done that a few times before.
... I did run the computer using the igp while the graphics card was out, but I didn't think that should cause a problem.
Anyway, I wiped the disk and installed everything again and it worked
Phew! Saves having to replace the motherboard.
I'll just be a bit more careful about putting in the graphics card with the Musashi cooler. I probably won't run the computer with igp this time while the graphics card it out.
In the meantime, I've had a little fun doing some overclocking.
http://www.overclock.net/t/1291703/ivy- ... therboards
http://rog.asus.com/129672012/maximus-v ... ing-guide/
http://www.overclock.net/t/1247413/ivy- ... at-the-end
Managed to get 4.5GHz stable while running Prime95. This is while keeping the top and front fans at a speed of around 400rpm. The cpu fan was around 1050rpm, which is inaudible against the sound of the top fan.
Asus support said to install everything again from a clean disk.
I did uninstall the drivers and deleted the directories used, but that didn't work.
In theory, taking a graphics card out and plugging it back in again should cause any problems. I've done that a few times before.
... I did run the computer using the igp while the graphics card was out, but I didn't think that should cause a problem.
Anyway, I wiped the disk and installed everything again and it worked
Phew! Saves having to replace the motherboard.
I'll just be a bit more careful about putting in the graphics card with the Musashi cooler. I probably won't run the computer with igp this time while the graphics card it out.
In the meantime, I've had a little fun doing some overclocking.
http://www.overclock.net/t/1291703/ivy- ... therboards
http://rog.asus.com/129672012/maximus-v ... ing-guide/
http://www.overclock.net/t/1247413/ivy- ... at-the-end
Managed to get 4.5GHz stable while running Prime95. This is while keeping the top and front fans at a speed of around 400rpm. The cpu fan was around 1050rpm, which is inaudible against the sound of the top fan.